Ringing In The Rain:

An Agent Based Weather Warning System

Introduction

People’s daily lives are impacted by lots of dynamic environmental information, such as weather and traffic. Though most of this information is available on the Internet now, there is no easy way for someone to access that kind of information while in a mobile state. In addition, people do not need to know this kind of information all the time unless there is some significant change that may impact their current or future activities.

Ringing In The Rain proposes creating a distributed multi-agent architecture that uses GPS-enabled cell phones to build a mobile service development framework. The goal of this framework is to build mobile services to deliver timely changes in environmental information that could impact a user’s current or future activities.

Problem Scenario

In many countries, the population of bicycle riders has surged during the past few years for various reasons, such as the cost of gas, traffic control regulation, health issues, and leisure. Regardless of the reason people ride bicycles, one of the most frustrating things for them is being caught in rain that they could have avoided by taking a detour, if they had known they were heading toward a storm beforehand as Figure 1 illustrated.

This frustration leads to a bigger and more general issue: is there any prompt and precise way to offer people information which may impact their current or future activities while they are in a state of mobility with limited information accessibility?

Figure 1. Scenario of Riding Toward a Storm

Proposed Solution

To solve the aforementioned problem, offering people information may impact their current and future activities promptly and precisely while they are in a mobile state with limited information accessibility, Ringing In The Rain plans to combine the distributed multi-agent architecture with GPS-enabled cell phones to create an application framework that enables the creation of various mobile services delivering such information at the right time and place.

This framework will define several kinds of software agents, including 1) Data collection agent, 2) User-movement estimation agent, 3) Personal information management agent, and 4) User goal-maintaining agent, as shown in Figure 2.

Figure 2. The Architecture of Weather Warning System Based on Proposed Distributed Multi-Agent Framework

User movement estimation agent runs on GPS-enabled cell phones utilize GPS as location sensor. By collecting users’ movement data, User movement estimation agent can build users’ movement model and estimate users’ movement accordingly. Figure 3 shows the home/office commutation movement model of one of the Ringing In The Rain trial users.

Figure 3. User Movement Model from Collected Movement Data

Data collection agent collects various dynamic environmental data from Internet that may impact users. Personal information management agent provides users’ schedule for better movement estimation. User goal-maintaining agent processes data offered by other 3 kinds of agents and determines the information to send to users.

Demonstration System

A weather warning system for bicycle riders is built based on proposed framework to demonstrate its utility. With User movement estimation agent and Personal information management agent, this weather warning system tracks and estimates bicycle riders’ location as well as current weather patterns, monitoring by Data collection agent; User goal-maintaining agent, which tries to keep users dry, will warn riders about the risk of getting caught in the rain if it estimates that riders are heading into areas where rain may occur. With this system, riders can make better-informed decisions about which routes to take to avoid encounters with nasty weather conditions. Figure 4 to 7 show the user interface of this system and information users may get.

Figure 4. May rain while traveling at known route.	Figure 5. May rain while staying at known location.
Figure 6. May rain while traveling at unknown route.	Figure 7. May rain while staying at unknown location.

Contact

Student: Chaochi Chang (ccchang at media dot mit dot edu)

Advisor: Chris Schmandt (geek at media dot mit dot edu)